Loading…
Loading…
What is a skill?
A skill is markdown that the AI Delivery Chain ingests at run time. Skills are authored once and reused across many engagements.
Steer how the AI generates: outline overrides, prescribed sections, mandatory inclusions, formatting rules. Bind generation to the consultant's playbook.
Curated background knowledge. best practices, prior decisions, vendor docs, gotchas. Compressed into context the AI uses across the chain run.
How skills steer the AI
Skills dictate outline, conjure deployment files, anchor the model to curated references, and bind generation to a known-good playbook.
An instruction skill can declare its own outline. The AI Delivery Chain emits exactly those sections instead of the default. useful for industry-specific BRDs or proposal structures with mandatory clauses.
A `manifest:scripts` directive prompts the AI to enumerate every deployment file the engagement needs. Each file becomes its own per-file generation iteration with its own deliverable row.
`file:<path>` and `script:<name>` directives produce separate `type='script'` deliverables. NetSuite SuiteScripts, configurations, integration templates. alongside the prose artifacts.
Skills can attach reference URLs. When grounding is on, the chain fetches each URL via a curated proxy and inlines the content as preferred source material. keeping output anchored to known-good references.
Guided Skill Builder
The Guided Skill Builder walks consultants through publishing in 10 steps. Reference skills skip the generation-control steps.
Start from a 0a generic, 0b lean enhancement, or 1a IC example template.
Instruction (steers generation) or reference (background expertise).
Title, summary, platform, modules, industries, tags, visibility.
(Instruction only) Override the chain's default outline with your own.
(Instruction only) Attach URLs that the chain fetches via Jina at run time.
(Instruction only) Declare manifest:scripts or file:<path> directives for per-file generation.
Write the actual skill content with the editor's live preview.
Free, per-run, or revenue share. Tier-bound price bands enforced by trigger.
Dry-run the skill against a sample requirement; review the generated artifact.
Submit for moderation; bronze tier on first publish; tiers rebalance with use + ratings.
Templates available today: three 0a. generic starters · three 0b. lean enhancement examples · three 1a. IC examples. New templates ship as the marketplace matures.
Quality tiers
Tiers are computed from usage, AM ratings, and dispute history. They live inside the product where AMs rate and consultants iterate.
Default tier
Every published skill starts here. Visible to Account Managers, eligible for use in the AI Delivery Chain.
5 uses · 4.0+ avg rating
Skill has been used in at least five chain runs and holds an average rating of 4.0 or above from Account Managers.
15 uses · 4.3+ avg · 3+ reviews
Battle-tested skill with strong AM ratings and qualitative review depth.
50 uses · 4.5+ avg · 10+ reviews · 0 disputes
The top-tier skills in the marketplace. Reliably drive high-quality artifacts across engagements with no open disputes.
Tiers are an in-product signal. Specific consultant rankings and per-skill ratings live inside the workspace for Account Managers and authors. they're not published on the public site.
Leaderboard preview
Composite score = (gold + platinum count) × avg rating × total uses. Quarterly + all-time. Public site shows shape only. actual identities live behind app login.
Illustrative. Real leaderboard data is in-product only.
Buyers don't pick skills directly. That's the AM's job. But the skill marketplace is why your delivery package runs on real consultant playbooks, not generic AI output. Every skill that drove your engagement appears in the audit trail with a per-skill Value Analysis.
See the AI Delivery Chain